基于小波变换的数字图像自适应盲水印算法及实现  

Adaptive Digital Image Blind Watermarking Algorithm Based DWT and Its Realization

作　　者：王智文[1] 刘美珍[2] 蔡启先[1] 李道丰[1] 谢国庆[3] 

机构地区：[1]广西工学院计算机工程系,广西柳州545006 [2]广西工学院图书馆,广西柳州545006 [3]福建师范大学软件学院,福建福州350007

出　　处：《计算机测量与控制》2009年第11期2273-2275,2279,共4页Computer Measurement &Control

基　　金：广西教育厅科研项目(200707LX196;200707LX182);广西工学院自然科学基金项目(院科自0704102)

摘　　要：针对传统基于人类视觉系统特性的空间域与变换域结合算法的缺陷,提出了一种基于小波变换的数字图像自适应盲水印新算法,通过置乱加密技术提高整个数字水印系统的安全性能,水印嵌入过程中使用了改进的量化水印算法;通过计算相邻特征平均值来自适应调整待量化小波系数的量化间隔,大大提升了水印系统的逼真度;实验结果表明,该算法具有嵌入信息量大,嵌入水印图像透明性好,对常见的噪声、中值滤波、旋转、剪切和JPEG压缩抵抗性强等特点,算法对于常见攻击后提取出的水印仍与原始水印具有0.90以上的相似度,具有很好的鲁棒性。A novel self adaptive digital image blind watermarking algorithm is put forward without using the traditional method of combining spatial and DWT domain based on the characteristics of human visual system. The security capabilities of numeric watermarking system are improved by using chaos encryption techniques. The scheme can adaptively adjust the quantized interval by calculating the border upon characters＇ mean value of the wavelet coefficients which is selected to embed watermark. The fidelity of digital image blind watermarking system is upgraded by this method. The experimental results indicate that the proposed algorithm has the following characters： large embed information capacity, good invisibility of embed watermarking image and robustness against attacks such as additive yawp, median filtering, circumrotation, cutting, JPEG compression, etc. The similarity between the original watermark and extracted watermark is above 0.90.

关 键 词：盲水印 自适应量化 离散小波变换 鲁棒性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]